Impact of complexity on daily and multi-step forecasting of streamflow with chaotic, stochastic, and black-box models

Tongal, Hakan; Berndtsson, Ronny

doi:10.1007/s00477-016-1236-4

Cited by 32 publications

(17 citation statements)

References 99 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…If the calculated c e is unchanged by increasing the number of embedding dimensions, C e can be considered as the correlation dimension of the attractor in the system. But, if c e is not stable as a function of embedding dimensions, this system can be considered non-chaotic [54,96,97].…”

Section: Correlation Dimensionmentioning

confidence: 99%

Nonlinear Dynamic Modeling of Urban Water Consumption Using Chaotic Approach (Case Study: City of Kelowna)

Yousefi

Courtice

Naser

et al. 2020

Water

View full text Add to dashboard Cite

This study investigated urban water consumption complexity using chaos theory to improve forecasting performance to help optimize system management, reduce costs and improve reliability. The objectives of this study were to (1) investigate urban water distribution consumption complexity and its role in forecasting technique performance, (2) evaluate forecasting models by periodicity and lead time, and (3) propose a suitable forecasting technique based on operator applications and performance through various time scales. An urban consumption dataset obtained from the City of Kelowna (British Columbia, Canada) was used as a test case to forecast future consumption values using varying lead times under different temporal scales to identify models which may improve forecasting performance. Chaos theory techniques were employed to inform model optimization. This study attempted to address the paucity of studies on chaos theory applications in water consumption forecasting. This was accomplished by applying non-linear approximation, dynamic investigation, and phase space reconstruction for input variables, to improve the accuracy in various periodicity and lead time. To reconstruct the phase space, lag time was calculated using average mutual information for daily resolution as 17 days to reconstruct the phase space. The optimum embedding dimension and correlation exponent for the phase space were 18 and 3.5, respectively. Comparing the results, the non-linear local approximation model provided the best performance. The forecasting horizon for the models was 122 days. Moreover, phase space reconstruction improved the accuracy of the models for the different lead times. The findings of this study may improve forecasting performance and provide evidence to support further investigation of the chaotic behaviour of water consumption values over different time scales.

show abstract

Section: Correlation Dimensionmentioning

confidence: 99%

Nonlinear Dynamic Modeling of Urban Water Consumption Using Chaotic Approach (Case Study: City of Kelowna)

Yousefi

Courtice

Naser

et al. 2020

Water

View full text Add to dashboard Cite

show abstract

“…Furthermore, Hu et al (2001) and Tongal and Berndtsson (2016) perform multi-step ahead forecasting of time series without seasonal behaviour.…”

Section: Time Series Forecasting In Hydrology and Beyondmentioning

confidence: 99%

“…Similarly, Yu et al (2004) compare several forecasting methods, including an ARIMA model and SVM, on two daily time series of runoff and Tongal and Berndtsson (2016) compare several stochastic and ML forecasting methods on three time series of streamflow processes. Additionally, in Chen et al (2012) the reader can find one of the few studies using RF for hydrological forecasting tasks.…”

Section: Right After the Introduction Of The Currently Classical Automentioning

confidence: 99%

Comparison of Stochastic and Machine Learning Methods for Multi-Step Ahead Forecasting of Hydrological Processes

Papacharalampous¹,

Tyralis²,

Koutsoyiannis³

2017

Preprint

View full text Add to dashboard Cite

Abstract:We perform an extensive comparison between 11 stochastic to 9 machine learning methods regarding their multi-step ahead forecasting properties by conducting 12 large-scale computational experiments. Each of these experiments uses 2 000 time series generated by linear stationary stochastic processes. We conduct each simulation experiment twice; the first time using time series of 110 values and the second time using time series of 310 values. Additionally, we conduct 92 real-world case studies using mean monthly time series of streamflow and particularly focus on one of them to reinforce the findings and highlight important facts. We quantify the performance of the methods using 18 metrics. The results indicate that the machine learning methods do not differ dramatically from the stochastic, while none of the methods under comparison is uniformly better or worse than the rest. However, there are methods that are regularly better or worse than others according to specific metrics.

show abstract

“…Studies have also found that AR models were quite competitive with the complex nonlinear models including k-nearest neighbor and ANN models. (Tongal and Berndtsson 2016). In this regard, the significant flow persistence represents an important feature in flood forecasting and the AR(2) model is simple enough, while capturing the flow persistence, to suffice a bench mark series.…”

Section: Uncertainties In Mpe Criteriamentioning

confidence: 99%

On the criteria of model performance evaluation for real-time flood forecasting

Cheng

Lien²,

et al. 2016

Stoch Environ Res Risk Assess

View full text Add to dashboard Cite

Model performance evaluation for real-time flood forecasting has been conducted using various criteria. Although the coefficient of efficiency (CE) is most widely used, we demonstrate that a model achieving good model efficiency may actually be inferior to the naïve (or persistence) forecasting, if the flow series has a high lag-1 autocorrelation coefficient. We derived sample-dependent and AR model-dependent asymptotic relationships between the coefficient of efficiency and the coefficient of persistence (CP) which form the basis of a proposed CE-CP coupled model performance evaluation criterion. Considering the flow persistence and the model simplicity, the AR(2) model is suggested to be the benchmark model for performance evaluation of real-time flood forecasting models. We emphasize that performance evaluation of flood forecasting models using the proposed CE-CP coupled criterion should be carried out with respect to individual flood events. A single CE or CP value derived from a multi-event artifactual series by no means provides a multi-event overall evaluation and may actually disguise the real capability of the proposed model.

show abstract

Impact of complexity on daily and multi-step forecasting of streamflow with chaotic, stochastic, and black-box models

Cited by 32 publications

References 99 publications

Nonlinear Dynamic Modeling of Urban Water Consumption Using Chaotic Approach (Case Study: City of Kelowna)

Nonlinear Dynamic Modeling of Urban Water Consumption Using Chaotic Approach (Case Study: City of Kelowna)

Comparison of Stochastic and Machine Learning Methods for Multi-Step Ahead Forecasting of Hydrological Processes

On the criteria of model performance evaluation for real-time flood forecasting

Contact Info

Product

Resources

About